CN110135282B - Examinee return plagiarism cheating detection method based on deep convolutional neural network model - Google Patents
Examinee return plagiarism cheating detection method based on deep convolutional neural network model Download PDFInfo
- Publication number
- CN110135282B CN110135282B CN201910336886.XA CN201910336886A CN110135282B CN 110135282 B CN110135282 B CN 110135282B CN 201910336886 A CN201910336886 A CN 201910336886A CN 110135282 B CN110135282 B CN 110135282B
- Authority
- CN
- China
- Prior art keywords
- cheating
- examinee
- head
- plagiarism
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an examinee's back-heading plagiarism cheating detection method based on a deep convolutional neural network model, which detects the positions of examinee's joints and limbs by adopting a VGG deep neural network model; calculating the central point p of five key points of headc(x, y) calculating a head center point p using a distance metric criterionc(x, y) and shoulder keypoints psDistance d between (x, y)(c,s)To judge whether the suspected cheating condition is satisfied; passing the center point p of the frame header of the previous framec f(x, y) and the current frame head center point pc bDifference d in displacement between (x, y)(f,b)And returning times to judge whether the cheating conditions are met; and then, whether the examinee has the behavior of plagiarism cheating by returning is judged by further calculating the confidence difference of key points on two sides of the head. According to the method, the suspected back-plagiarism cheating examinees in the video are detected, the image information of cheating behaviors of the examinees can be accurately extracted and positioned, the workload of related personnel is greatly reduced, and the working efficiency and the accuracy are improved.
Description
Technical Field
The invention relates to the technical field of computer vision and pattern recognition, in particular to a test taker return plagiarism cheating detection method based on a deep convolutional neural network model.
Background
The system or the method for detecting cheating behaviors of the examinees becomes a new application direction in the fields of digital image processing, pattern recognition and computer vision analysis, and the research on the detection of the cheating behaviors of the examinees is a key technology for intelligent information popularization and construction of various education culture examination fields in the future. Has certain leading edge and guiding function, and has great market application value and social significance.
The research of the test method for the cheating behaviors of the examinees is a big problem in the field of computer vision at present, and at present, many scholars at home and abroad carry out deep research around deep learning neural networks, but the cases of applying the technology to the field of the cheating behaviors of the examinees are few.
The patent "a based on big data analysis online examination invigilates system" (CN201711022450.0), regard network resources such as customer end, Web server, cloud storage as the carrier, through whether opening etc. to IP address, gateway information, camera, judge whether the examinee has the online behavior of cheating. The patent "a management system of practising fraud is prevented in examination room" (CN201810152948.7), realizes the upgrading of hardware condition based on internet of things, through a plurality of image entry modules of installing in the classroom, realizes reminding in real time to examinee's action. The patent "method for carrying out intelligent video identification to examination cheating incident" (CN201010226332.3), adopts the gaussian mixture model to separate the foreground image and the background image in the video, trains the decision tree with the motion pixel characteristic value as the input data, and the method uses single color information as the classification characteristic, belongs to the shallow layer characteristic information expression, and does not fully consider the depth characteristic information such as texture, edge, space, etc. In addition, the method takes the whole state of the examinee as the classification information, loses the characteristics of the local part of the examinee, and cannot give sufficient expression to the characteristic difference of the cheating examinee and the non-cheating examinee. The depth features contain a large amount of feature information and can be effectively used for analyzing and identifying tiny movements of joints of examinees. In addition, there are some researches on a method for detecting cheating on various objects, such as "a method and apparatus for detecting cheating sites" (CN201711310287.8), a method for detecting whether a cheating operation is performed on a website by collecting information on the ratio of vibration of a site, "a cheating detection system in a casino" (CN201680045982.3), a cheating detection system for detecting cheating in a game in which chips are collected and paid back according to the result of winning or losing, and the like. However, neither such cheating detection system nor detection method can introduce therein a deep neural network technology, which is one of the research hotspots in the field of computer vision at present, and the application of the technology in this field is still blank.
Disclosure of Invention
The technical task of the invention is to provide a test method for the cheating of the back-heading copying of the examinee based on the deep convolutional neural network model, which is used for supporting the deep convolutional neural network, determining the position and the confidence coefficient of a key point by extracting the deep characteristic information of the examinee, then adopting a certain rule to carry out hierarchical judgment on the examinee suspected to have the back-heading cheating behavior, and finally making the final judgment on whether the examinee cheats according to the confidence coefficient.
The technical scheme adopted by the invention for solving the technical problems is as follows: an examinee's back-heading plagiarism cheating detection method based on a deep convolutional neural network model comprises the following steps:
and 4, if the secondary judgment result also meets the cheating condition, further calculating the confidence difference of key points on two sides of the head, if the difference value is not less than a threshold value, making final judgment, and judging that the examinee has the behavior of copying and cheating by returning the head.
Compared with the prior art, the technical scheme of the invention has the advantages that:
1. according to the method, the behavior of the examinees who are suspected to cheat by plagiarism is detected by a deep neural network model feature extraction method based on computer vision, and the image information of the cheating behavior of the examinees can be accurately extracted and positioned by detecting the examinees suspected to cheat by plagiarism in the video, so that the workload of related personnel is greatly reduced, the working efficiency and the accuracy rate are improved, and the method has application value and popularization prospect;
2. the examinee sitting posture turning prediction method based on the key point technology is provided, the central point information of a plurality of key points of the head is calculated at the same time, and a turning judgment rule is formed by combining the shoulder position information;
3. the method is used for judging whether back cheating exists or not according to the confidence difference characteristics of symmetrical parts on two sides of the head of the examinee.
Drawings
FIG. 1 is a schematic flow diagram of a process of the present invention;
FIG. 2 is a distribution diagram of key points of a human body;
FIG. 3 is a graph of key point detection results;
FIG. 4 is a graph of confidence calculation results;
fig. 5 is a graph of the cheating detection results.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The invention provides a test taker return plagiarism cheating detection method based on a deep convolutional neural network model, the flow is shown in figure 1, and the specific technical scheme is as follows:
firstly, to finish the detection work of 18 key points of the whole body of the examinee in the image and obtain the specific coordinate position of each key, the distribution condition of 18 key points of the human body is shown in fig. 2, wherein the key points of the head are distributed most densely, and necessary support is provided for accurately analyzing the direction and the displacement of the head movement of the examinee;
in order to extract depth feature information of an image, effectively and accurately express key points of each part of an examinee and realize detection of the key points and calculation of confidence coefficients, the invention adopts a classic VGG (visual Geometry Group Net) depth convolution neural network to extract the feature information of each part of the examinee to obtain a feature vector set F, completes training of each model, and obtains a corresponding key point coordinate set P and a confidence coefficient C corresponding to each key point by calculating an integral graph of each joint part, as shown in a formula (1) and a formula (2):
P＝ρ(F) (1)
wherein, P ═ { P ═ P(1),p(2),…,p(n)The method comprises the steps that key point information of n examinees in an image is represented, G represents the actual position of the key point in the image, rho represents a convolutional neural network processing process, sigma is a parameter of Gaussian distribution, and C is { C ═ C }0,c1,…,c17}; the VGG deep convolution neural network that the extraction image characteristic adopted has 13 layers in total, including convolution layer 10 layers, pooling layer 3 layers, and the size of convolution layer all sets up to: 3 × 3, the sizes of all the pooling layers are set to be 2 × 2; the results of keypoint detection for a test taker and their corresponding confidence level are shown in fig. 3 and 4.
Secondly, after the preprocessing stage of key point information acquisition is completed, independent preliminary judgment can be carried out on whether each examinee has suspected back cheating behaviors, and because the method aims at detecting the back cheating of the examinee, only partial key points of the head and the shoulders of the body are predicted and judgedc(x, y), the calculation method is shown in formula (3):
pc(x,y)＝((pmin(x)+pmax(x))/2,(pmin(y)+pmax(y))/2) (3)
wherein p ismax(x) And pmin(x) Respectively represent the maximum value and the minimum value in the abscissa of five key points of the head, namely,
pmax(y) and pmin(y) respectively represents the maximum value and the minimum value in the vertical coordinates of the five key points of the head,
k is the subscript of the keypoint. Then, based on the obtained head center point position and the coordinate information of the examinee's shoulder key points obtained in the first step, the position relationship between the head and the shoulder can be calculated, and the distance measurement criterion used here is shown in formula (4):
d(c,s)＝|pc(x)-ps(x)| (4)
wherein p iss(x) Is the abscissa value of the shoulder key point, s takes the value of 2 or 5 to represent the right shoulder or the left shoulder, d(c,s)When the distance between the head center point and the shoulders is smaller than 1/3 of the distance between the shoulders of the examinee, the examinee can judge that the action of the suspected return occurs, the continuous N-frame images are taken as a sliding window, and when the number of times of the suspected return action in the window exceeds a threshold value J1Then, it can be preliminarily determined that the examinee has suspected plagiarism cheating behavior.
Thirdly, judging the examinee suspected of cheating again, taking one frame of image every four frames in the current sliding window, calculating the displacement relation between the central points in the two frames by taking the examinee head central point obtained in the second step as an object in the front and back adjacent frames in all the images, and further obtaining the direction of turning the head, wherein the method is specifically shown in a formula (5):
wherein d is(f,b)Representing the displacement of the examinee's head in two successive frames, pc b(x, y) is the coordinate of the center point of the examinee's head in the current frame, pc f(x, y) is the coordinate of the head center point of the examinee in the previous frame, if the linear displacement exceeds a certain size, the examinee is judged to have the turn-back action, if the examinee is in a sliding window and the turn-back action occurrence frequency in the same direction exceeds a threshold value J2Then, the examinee can be further judged to be suspected of plagiarism cheating again.
The fourth step is that the first step is that,finally judging the image frames with suspected cheating back in the second step and the third step, and according to the confidence values c of the eyes and the ears of the examinees obtained in the first step14、c15、c16And c17Calculating the confidence difference between the left eye and the right eye and the confidence difference between the left ear and the right ear of the image frames containing the suspected return motion detected in the second step and the third step respectively, and if the difference of any part is not lower than a threshold J3That is to say that the first and second electrodes,
then, it can be determined that the frame belongs to the back cheating action, after all the suspected frames are determined one by one, if the reserved number is still not lower than the respective threshold value, J is determined1And J2And then, final judgment can be made, and the examinee is judged to have the behavior of copying and cheating.
Threshold value J in the present invention1、J2And J3Can be set manually according to actual requirements.
The back cheating detection result of the method is shown in fig. 5, when the position of the center point of the head of the examinee, which is close to the shoulder, is relatively large in displacement of frames before and after the head turning action and relatively obvious in confidence difference value of the left side and the right side of the head, the examinee can be judged to have the back cheating behavior, and experiments prove that the detection effect is relatively good.
The technical idea of the present invention is described in the above technical solutions, and the protection scope of the present invention is not limited thereto, and any changes and modifications made to the above technical solutions according to the technical essence of the present invention belong to the protection scope of the technical solutions of the present invention.
Claims (6)
1. A deep convolutional neural network model-based examinee return plagiarism cheating detection method is characterized by comprising the following steps:
step 1, extracting the features of an image by adopting a VGG deep neural network model to obtain a feature vector F, detecting key points of each part of an examinee in the image to obtain a set P, and calculating a corresponding confidence coefficient C;
step 2, calculating the central points p of five key points at the headc(x, y) finding the head center point p using a distance metric criterionc(x, y) and shoulder keypoints psDistance d between (x, y)(c,s)And judging whether a suspected return plagiarism cheating condition is met, wherein the abscissa value p of the shoulder key points(x) S is 2 or 5, which represents the distance d between the right shoulder and the left shoulder(c,s)When the distance between shoulders of the examinee is less than 1/3, the examinee can be judged to have the suspected return behavior, the continuous N-frame images are taken as a sliding window, and when the number of suspected return actions in the window exceeds a threshold value J1If yes, the examinee can be preliminarily judged to have suspected plagiarism cheating behavior;
and 3, if the suspected plagiarism cheat of returning the head is judged to exist, further secondary judgment is carried out, one frame of image is taken from the current sliding window every other four frames, and in the front and back adjacent two frames of all the frame images taken from every other four frames, the central point p in the head of the front frame in the adjacent two frames of images is calculatedc f(x, y) and the current frame head center point pc bDifference d in displacement between (x, y)(f,b)And returning times, and judging whether the conditions of returning, copying and cheating are met, wherein if the displacement difference d(f,b)If the number of times of the turn-back actions in the same direction exceeds a threshold value J in a sliding window, the turn-back action is judged to exist2Then, the suspicion that the examinee returns to the cheating behavior of plagiarism can be further judged;
and 4, if the secondary judgment result also meets the condition of back-heading copying cheating, further calculating the confidence difference of key points on two sides of the head, if the difference is not lower than a threshold value, making final judgment, and judging that the examinee has the back-heading copying cheating behavior.
2. The method for detecting the cheating of the back-heading plagiarism of the test taker based on the deep convolutional neural network model according to claim 1, wherein the step 1 specifically comprises the following steps:
extracting feature information of each part of the examinee by adopting a VGG deep convolution neural network to obtain a feature vector set F, completing iterative training of a network model, and obtaining a corresponding key point coordinate set P and a confidence coefficient C corresponding to each key point by calculating an integral map of each joint part, as shown in a formula (1) and a formula (2):
P＝ρ(F) (1)
wherein, P ═ { P (1), P (2), …, P (n) } represents key point information of n examinees in the image, G represents the actual position of the key point in the image, ρ represents the convolutional neural network processing procedure, σ is a parameter of gaussian distribution, and C ═ { C ═ C0,c1,…,c17}。
3. The method for detecting the cheating of the back-heading and plagiarism of the examinee based on the deep convolutional neural network model as claimed in claim 2, wherein the VGG deep convolutional neural network adopted for extracting the image features in the step 1 has 13 layers, wherein the VGG deep convolutional neural network comprises 10 layers of convolutional layers and 3 layers of pooling layers, and the sizes of the convolutional layers are all set as: 3 × 3, the sizes of the pooling layers are all set to 2 × 2.
4. The method for detecting the cheating of the back-heading plagiarism of the examinee based on the deep convolutional neural network model of claim 1, wherein the head center point p in the step 2c(x, y), and a head center point pc(x, y) and shoulder keypoints psDistance d between (x, y)(c,s)The calculation method comprises the following steps:
head center point pcThe calculation method of (x, y) is shown in formula (3):
pc(x,y)＝((pmin(x)+pmax(x))/2,(pmin(y)+pmax(y))/2) (3)
wherein p ismax(x) And pmin(x) Respectively represent the maximum value and the minimum value in the abscissa of five key points of the head, namely,
pmax(y) and pmin(y) respectively represents the maximum value and the minimum value in the vertical coordinates of the five key points of the head,
distance d between head center point and shoulder(c,s)The distance metric criterion used for the calculation of (2) is shown in equation (4):
d(c,s)＝|pc(x)-ps(x)| (4)
wherein p isc(x) Is the abscissa of the center point of the head, ps(x) The abscissa value of the shoulder key point.
5. The method for detecting the back-heading plagiarism cheating of the examinee based on the deep convolutional neural network model according to claim 1, wherein the method for judging whether the back-heading plagiarism cheating condition is met in the step 3 is as follows:
taking one frame of image every other four frames from the current sliding window, taking the central point of the head of the examinee obtained in the step 2 as an object in the two adjacent frames in the front and back of all the frame images taken out every other four frames, calculating the displacement relation between the central points in the two frames to further obtain the direction of the turning head, and calculating the displacement difference of the head of the examinee according to a formula (5):
wherein d is(f,b)Representing the displacement difference, p, of the examinee's head in the two previous and subsequent framesc b(x, y) is the coordinate of the center point of the examinee's head in the current frame, pc f(x, y) is preThe head center point coordinates of the examinee in one frame.
6. The method for detecting the cheating of the back-heading plagiarism of the test taker based on the deep convolutional neural network model as claimed in claim 5, wherein the step 4 specifically comprises the following steps:
finally judging the image frames with suspected back cheating detected in the steps 2 and 3 according to the confidence values c of the eyes and the ears of the examinees obtained in the step 114、c15、c16And c17Calculating the confidence difference between the left eye and the right eye and the confidence difference between the left ear and the right ear of the image frames containing the suspected return motion detected in the step 2 and the step 3 respectively, and if the difference of any part is not lower than a threshold J3That is to say that the first and second electrodes,
then, it can be determined that the frame belongs to the back cheating action, after all the suspected frames are determined one by one, if the reserved number is still not lower than the respective threshold value, J is determined1And J2And then, final judgment can be made, and the examinee is judged to have the behavior of copying and cheating.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910336886.XA CN110135282B (en) | 2019-04-25 | 2019-04-25 | Examinee return plagiarism cheating detection method based on deep convolutional neural network model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910336886.XA CN110135282B (en) | 2019-04-25 | 2019-04-25 | Examinee return plagiarism cheating detection method based on deep convolutional neural network model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110135282A CN110135282A (en) | 2019-08-16 |
| CN110135282B true CN110135282B (en) | 2021-06-04 |
Family
ID=67570922
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910336886.XA Active CN110135282B (en) | 2019-04-25 | 2019-04-25 | Examinee return plagiarism cheating detection method based on deep convolutional neural network model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110135282B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110718067A (en) * | 2019-09-23 | 2020-01-21 | 浙江大华技术股份有限公司 | Violation behavior warning method and related device |
| CN110781763B (en) * | 2019-09-30 | 2022-06-17 | 沈阳图为科技有限公司 | Human body looking-at motion detection method based on posture |
| CN110837784B (en) * | 2019-10-23 | 2023-06-20 | 中山大学 | Examination room peeping and cheating detection system based on human head characteristics |
| CN112819021B (en) * | 2019-11-15 | 2024-05-07 | 北京地平线机器人技术研发有限公司 | Image detection method and device, electronic equipment and storage medium |
| CN113095968A (en) * | 2019-12-23 | 2021-07-09 | 深圳云天励飞技术有限公司 | Cheating identification method and device and electronic equipment |
| CN112232270A (en) * | 2020-10-29 | 2021-01-15 | 广西科技大学 | MDSSD face detection method based on model quantization |
| CN112686311A (en) * | 2020-12-30 | 2021-04-20 | 杭州电子科技大学 | Online examination system |
| CN114882533A (en) * | 2022-05-30 | 2022-08-09 | 北京百度网讯科技有限公司 | Examination room abnormal behavior detection method, device, equipment and storage medium |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102014274A (en) * | 2010-11-10 | 2011-04-13 | 深圳市海云天科技股份有限公司 | Examination monitoring system and monitoring information processing method thereof |
| CN102348101A (en) * | 2010-07-30 | 2012-02-08 | 深圳市先进智能技术研究所 | Examination room intelligence monitoring system and method thereof |
| CN102880781A (en) * | 2012-06-15 | 2013-01-16 | 北京理工大学 | On-line examination intelligent monitoring system |
| CN103136510A (en) * | 2013-01-17 | 2013-06-05 | 京东方科技集团股份有限公司 | Examination monitoring system and method |
| CN104378549A (en) * | 2014-10-30 | 2015-02-25 | 东莞宇龙通信科技有限公司 | Snapshot method and device and terminal |
| CN105959624A (en) * | 2016-05-03 | 2016-09-21 | 方筠捷 | Examination room monitoring data processing method and automatic monitoring system thereof |
| CN106791633A (en) * | 2016-12-14 | 2017-05-31 | 北华航天工业学院 | Intelligence invigilator's system and method |
| CN107077796A (en) * | 2016-03-11 | 2017-08-18 | 深圳市大疆创新科技有限公司 | Method, system and the equipment of the anti-cheating of storage medium, network test |
| CN107134195A (en) * | 2017-06-26 | 2017-09-05 | 上海眼控科技股份有限公司 | It is a kind of to be used for the pulling over observing determination methods of intelligent Driving Test |
| CN107491717A (en) * | 2016-06-13 | 2017-12-19 | 科大讯飞股份有限公司 | The detection method that cheats at one's exam and device |
| CN109460727A (en) * | 2018-10-31 | 2019-03-12 | 中国矿业大学 | A kind of examination hall monitoring system and method based on Human bodys' response |
-
2019
- 2019-04-25 CN CN201910336886.XA patent/CN110135282B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102348101A (en) * | 2010-07-30 | 2012-02-08 | 深圳市先进智能技术研究所 | Examination room intelligence monitoring system and method thereof |
| CN102014274A (en) * | 2010-11-10 | 2011-04-13 | 深圳市海云天科技股份有限公司 | Examination monitoring system and monitoring information processing method thereof |
| CN102880781A (en) * | 2012-06-15 | 2013-01-16 | 北京理工大学 | On-line examination intelligent monitoring system |
| CN103136510A (en) * | 2013-01-17 | 2013-06-05 | 京东方科技集团股份有限公司 | Examination monitoring system and method |
| CN104378549A (en) * | 2014-10-30 | 2015-02-25 | 东莞宇龙通信科技有限公司 | Snapshot method and device and terminal |
| CN107077796A (en) * | 2016-03-11 | 2017-08-18 | 深圳市大疆创新科技有限公司 | Method, system and the equipment of the anti-cheating of storage medium, network test |
| CN105959624A (en) * | 2016-05-03 | 2016-09-21 | 方筠捷 | Examination room monitoring data processing method and automatic monitoring system thereof |
| CN107491717A (en) * | 2016-06-13 | 2017-12-19 | 科大讯飞股份有限公司 | The detection method that cheats at one's exam and device |
| CN106791633A (en) * | 2016-12-14 | 2017-05-31 | 北华航天工业学院 | Intelligence invigilator's system and method |
| CN107134195A (en) * | 2017-06-26 | 2017-09-05 | 上海眼控科技股份有限公司 | It is a kind of to be used for the pulling over observing determination methods of intelligent Driving Test |
| CN109460727A (en) * | 2018-10-31 | 2019-03-12 | 中国矿业大学 | A kind of examination hall monitoring system and method based on Human bodys' response |
Non-Patent Citations (1)
| Title |
|---|
| 基于3D卷积双路神经网络的考场行为异常识别;于明学 等;《北京电子科技学院学报》;20181231;第26卷(第4期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110135282A (en) | 2019-08-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110135282B (en) | Examinee return plagiarism cheating detection method based on deep convolutional neural network model | |
| WO2020151489A1 (en) | Living body detection method based on facial recognition, and electronic device and storage medium | |
| CN104050471B (en) | Natural scene character detection method and system | |
| CN110837768B (en) | Online detection and identification method for rare animal protection | |
| CN107506722A (en) | One kind is based on depth sparse convolution neutral net face emotion identification method | |
| CN109101865A (en) | A kind of recognition methods again of the pedestrian based on deep learning | |
| CN109670430A (en) | A kind of face vivo identification method of the multiple Classifiers Combination based on deep learning | |
| CN105160310A (en) | 3D (three-dimensional) convolutional neural network based human body behavior recognition method | |
| CN105740758A (en) | Internet video face recognition method based on deep learning | |
| CN109614488B (en) | Text classification and image recognition-based distribution network live working condition judgment method | |
| CN107330371A (en) | Acquisition methods, device and the storage device of the countenance of 3D facial models | |
| CN104850825A (en) | Facial image face score calculating method based on convolutional neural network | |
| CN109712127A (en) | A kind of electric transmission line fault detection method for patrolling video flowing for machine | |
| CN107808376A (en) | A kind of detection method of raising one's hand based on deep learning | |
| CN116052222A (en) | Cattle face recognition method for naturally collecting cattle face image | |
| CN112819015A (en) | Image quality evaluation method based on feature fusion | |
| CN106529441A (en) | Fuzzy boundary fragmentation-based depth motion map human body action recognition method | |
| CN113313107A (en) | Intelligent detection and identification method for multiple types of diseases on cable surface of cable-stayed bridge | |
| CN108830222A (en) | A kind of micro- expression recognition method based on informedness and representative Active Learning | |
| CN107633527A (en) | Target tracking method and device based on full convolutional neural networks | |
| CN109522865A (en) | A kind of characteristic weighing fusion face identification method based on deep neural network | |
| CN109472228A (en) | A kind of yawn detection method based on deep learning | |
| CN116758421A (en) | Remote sensing image directed target detection method based on weak supervised learning | |
| CN110232314A (en) | A kind of image pedestrian's detection method based on improved Hog feature combination neural network | |
| CN110443277A (en) | A small amount of sample classification method based on attention model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20201106 Address after: Room d09-629, international software park, No. 863-9, shangshengou village, Hunnan District, Shenyang City, Liaoning Province Applicant after: Shenyang Tuwei Technology Co., Ltd Address before: 110136, Liaoning, Shenyang, Shenbei New Area moral South Avenue No. 37 Applicant before: SHENYANG AEROSPACE University |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |